Thermal vias

Feb 28, 2012 22 Replies

Anyone have a pointer to a good white paper on the use of thermal vias to steer heat from devices on top of a board to a heat sink below?



Particularly if you used the thing to build a product that didn't burn up?



Thanks.


-- My liberal friends think I'm a conservative kook. My conservative friends think I'm a liberal kook. Why am I not happy that they have found common ground?



Tim Wescott, Communications, Control, Circuits & Software

formatting link


?

It's been a long while since I read it, but TI had an app note that included discussion of thermal vias:

TI Application Report SLMA002A -- NOVEMBER 1997 -- Revised JUNE 2006 PowerPAD(TM) Thermally Enhanced Package slma002a.pdf

Usually the part data sheet will tell you what-cha-got-ta do, regarding heat sinking on the opposite side.

If the part has a thermal slug, I put as many thermal vias under the part that I can fit in. Vias are cheap. The "newer" plated over vias (epoxy fill then a plating cap) are cool for this, so the solder doesn't leak through the little holes.

something like this?:

formatting link

-Lasse

Actually it helps if the solder can leak into the holes. You don't need to worry about the stencil anymore. Without solder leaking into the holes the amount of solder paste under the part can be very critical.

Failure does not prove something is impossible, failure simply indicates you are not using the right tools... nico@nctdevpuntnl (punt=.) --------------------------------------------------------------

We do thermal vias a lot. But it's a messy subject, so some details would be helpful.

What kind of part do you want to cool, and where will the heat ultimately go?

John Larkin, President Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation

FETs (and one linear regulator) in TO-252 packages on top of a board, heat sink on bottom. No more than 5W/package unless I've really screwed up the design -- so, enough that I need to pay attention, but not so much that I need a Super High Tech solution. The heat sink is large enough that it's not going to be a significant participant in the thermal resistance, assuming that we can get the heat out to it.

The more I look and think and calculate the more I think that the heat sink on the bottom is going to make this a fairly easy problem to solve.

I think (but am not yet sure) that if I surround the FETs with thermal vias down to the bottom of the board, and mount the board to the heat sink with a sil-pad or something better (and choice of that "something" is soon to be a separate thread) then I'll be A-OK.

My liberal friends think I'm a conservative kook. My conservative friends think I'm a liberal kook. Why am I not happy that they have found common ground? Tim Wescott, Communications, Control, Circuits & Software http://www.wescottdesign.com

d

considered something like this?

formatting link

09DG/HS406-ND/1625533

-Lasse

Just a hint:

formatting link

Failure does not prove something is impossible, failure simply indicates you are not using the right tools... nico@nctdevpuntnl (punt=.) --------------------------------------------------------------

Here's a TI app note with some charts of thermal vias vs Rjc

Cheers

formatting link

That would be great if there's some air flow on the top of the board.

John Larkin, President Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation

I can't count on air flow. I can count on a nice cool wall next to my board, and a customer expectation for a heat sink.

Tim Wescott Control system and signal processing consulting www.wescottdesign.com

If you had actually read the post, instead of only the title, you would have seen where. Blind as a bat ditz.

formatting link

You are truly dumb, since he already said that the heat sink goes on the bottom of the board.

You should be taken away from the computer until you learn how to actually READ a post, BEFORE you are permitted to post a reply to it.

Refer to the CRC handbook of common sense, and the CRC handbook of PCB substrate thermal sinking.

Ask the PCB maker what their thickest PTH is and their annular ring sizes. You also make sure to have tie points for the via at each layer the PCB has. They do not need to connect to anything (other vias). They assist in the sinking capacity. Other "rules" they may have for such larger vias as well. No masking, of course, and solder fill them.

If you fill the vias with solder, you'll get bumps on the bottom of the board. Then the heat sink won't sit flat on the pcb, which will increase the thermal resistance.

John Larkin, President Highland Technology Inc www.highlandtechnology.com jlarkin at highlandtechnology dot com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom timing and laser controllers Photonics and fiberoptic TTL data links VME analog, thermocouple, LVDT, synchro, tachometer Multichannel arbitrary waveform generators

formatting link

I was suggesting that the heat sinking might be done on top, which would save a heap of hassle. These surface-mount heat sinks are great. If there's no air flow, a little surface-mount sink on top wouldn't dissipate 5 watts. It has been confirmed that there is no air flow available.

John Larkin, President Highland Technology Inc www.highlandtechnology.com jlarkin at highlandtechnology dot com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom timing and laser controllers Photonics and fiberoptic TTL data links VME analog, thermocouple, LVDT, synchro, tachometer Multichannel arbitrary waveform generators

You obviously know nothing of the process. And apparently no powers of resolution either. If the sink is against the hole, it cannot "bump".

That has to be one of the oldest, most simple PTH issues, dealt with decades ago by most.

A phrenologist would have a field day with you, I'm sure.

Look at it this way: a via is useful as long as the copper plating in the hole conducts more heat than the hole insulates (the hole being a better insulator than the PCB material. So, one can do simple BOE calcs using (volume) thermal conductivity of copper, of the PCB, and of air. If the hole is too small, there is insufficient increase of (overall) thermal conductivity, and if the hole is too large one has added insulation. Work on a unit area basis and in practice use an array of holes. Better yet, mill out an area and fill that with a block of metal. If you need high thermal conductivity AND good electrical insulation, use black diamond.

Well, if too much solder leaks (and there is virtually NO control) one can get a nasty blob/bump of solder on the "back" side.

In my experience with QFN packages the holes are only filled for a little bit (maybe 10%). I've never seen solder coming out (I'm talking about production runs of hundreds of boards).

Failure does not prove something is impossible, failure simply indicates you are not using the right tools... nico@nctdevpuntnl (punt=.) --------------------------------------------------------------

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required